Theories of Addiction
Bruce K. Alexander
In many cases, addiction theorists have now progressed beyond stereotyped disease conceptions of alcoholism or the idea that narcotics are inherently addictive to anyone who uses them. The two major areas of addiction theory??"those concerning alcohol and narcotics??"have had a chance to merge, along with theorizing about overeating, smoking, and even running and interpersonal addictions.
Yet this new theoretical synthesis is less than meets the eye: It mainly recycles discredited notions while including piecemeal modifications that make the theories marginally more realistic in their descriptions of addictive behavior. These theories are described and evaluated in this chapter as they apply to all kinds of addictions.
They are organized into sections on genetic theories (inherited mechanisms that cause or predispose people to be addicted), metabolic theories (biological, cellular adaptation to chronic exposure to drugs), conditioning theories (built on the idea of the cumulative reinforcement from drugs or other activities), and adaptation theories (those exploring the social and psychological functions performed by drug effects).
While most addiction theorizing has been too unidimensional and mechanistic to begin to account for addictive behavior, adaptation theories have typically had a different limitation. They do often correctly focus on the way in which the addict's experience of a drug's effects fits into the person's psychological and environmental ecology.
In this way drugs are seen as a way to cope, however dysfunctionally, with personal and social needs and changing situational demands. Yet these adaptation models, while pointing in the right direction, fail because they do not directly explain the pharmacological role the substance plays in addiction.
They are often considered??"even by those who formulate them??"as adjuncts to biological models, as in the suggestion that the addict uses a substance to gain a specific effect until, inexorably and irrevocably, physiological processes take hold of the individual. At the same time their purview is not ambitious enough (not nearly so ambitious as that of some biological and conditioning models) to incorporate nonnarcotic or nondrug involvements.
They also miss the opportunity, readily available at the social-psychological level of analysis, to integrate individual and cultural experiences.
How Is Alcoholism Inherited?
Cigarette smoking, alcoholism, and overweight??"like divorce, child abuse and religion??"run in families. This addictive inheritance has been most studied in the case of alcoholism. Studies endeavoring to separate genetic from environmental factors, such as those in which adopted-away offspring of alcoholics were compared to adopted children with nonalcoholic biological parents, have claimed a three to four times greater alcoholism rate for those whose biologic parents were alcoholic (Goodwin et al. 1973).
Vaillant (l983) approvingly cited the Goodwin et al. and other research indicating genetic causality in alcoholism (see especially Vaillant and Milofsky 1982), but his own research did not support this conclusion (cf. Peele 1983a). In the inner-city sample that formed the basis for Vaillant's primary analysis, those with alcoholic relatives were between three and four times as likely to be alcoholic as those without alcoholic relatives.
Since these subjects were reared by their natural families, however, this finding does not distinguish effects of alcoholic environment from inherited dispositions. Vaillant did find that subjects with alcoholic relatives they did not live with were twice as likely to become alcoholic as subjects who had no alcoholic relatives at all.
Yet further nongenetic influences remain to be partialed out of Vaillant's results. The chief of these is ethnicity: Irish Americans in this Boston sample were seven times as likely to be alcohol dependent as were those of Mediterranean descent. Controlling for such large ethnicity effects would surely reduce the 2 to 1 ratio (for subjects with alcoholic relatives compared to those without) in alcoholism substantially even as other potential environmental factors that lead to alcoholism (besides ethnicity) would still remain to be controlled for.
Vaillant reported two other tests of genetic causality in his sample. He disconfirmed Goodwin's (1979) hypothesis that alcoholics with alcoholic relatives??"and hence a presumed inherited predisposition to alcoholism??"inevitably develop problems with drinking earlier than do others. Finally, Vaillant found no tendency for the choice of moderate drinking versus abstinence as a resolution for drinking problems to be related to number of alcoholic relatives, although it was associated with the drinker's ethnic group.
Proposing genetic mechanisms in alcoholism on the basis of concordance rates does not provide a model of addiction. What are these mechanisms through which alcoholism is inherited and translated into alcoholic behavior? Not only has no biological mechanism been found to date to underlie alcoholism, but research on alcoholics' behavior indicates that one cannot be found in the case of the loss of control of drinking that defines alcoholism.
Even the most severely alcoholic individuals "clearly demonstrate positive sources of control over drinking behavior" so that "extreme drunkenness cannot be accounted for on the basis of some internally located inability to stop" (Heather and Robertson 1981: 122). Intriguingly, controlled-drinking theorists like Heather and Robertson (1983) propose exceptions to their own analyses: Perhaps "some problem drinkers are born with a physiological abnormality, either genetically transmitted or as a result of intrauterine factors, which makes them react abnormally to alcohol from their first experience of it" (Heather and Robertson 1983: 141).
While it is certainly a fascinating possibility, no research of any type supports this suggestion. Vaillant (1983) found that self-reports by AA members that they immediately succumbed to alcoholism the first time they drank were false and that severe drinking problems developed over periods of years and decades.
The exceptions to this generalization were psychopaths whose drinking problems were components of overall abnormal lifestyles and behavior patterns from an early age. However, these kinds of alcoholics showed a greater tendency to outgrow alcoholism by moderating their drinking (Goodwin et al. 1971), indicating they also do not conform to a putative biological model. Prospective studies of those from alcoholic families also have failed to reveal early alcoholic drinking (Knop et al.1984).
Findings like these have led genetic theorists and researchers instead to propose that the inherited vulnerability to alcoholism takes the form of some probabilistically greater risk of developing drinking problems. In this view a genetic tendency??"such as one that dictates a drinker will have an overwhelming response to alcohol??"does not cause alcoholism. The emphasis is instead on such biological abnormalities as the inability to discriminate blood alcohol level (BAL), which leads alcoholics to show less effect from drinking and to drink more without sensing their condition (Goodwin 1980; Schuckit 1984).
Alternately, Schuckit (1984) proposed that alcoholics inherit a different style of metabolizing alcohol, such as producing higher levels of acetaldehyde due to drinking. Finally, Begleiter and other theorists have proposed that alcoholics have abnormal brain waves prior to ever having drunk or that drinking creates unusual brain activity for them (Pollock et al. 1984; Porjesz and Begleiter 1982).
All these theorists have indicated that their results are preliminary and require replication, particularly through prospective studies of people who become alcoholics. Negative evidence, however, is already available. Several studies have found that sensitivity to BAL, peak BAL after drinking, and elimination of blood alcohol are unrelated to family histories of alcoholism (Lipscomb and Nathan 1980; Pollock et al. 1984).
Other negative evidence for both BAL discrimination and metabolic hypotheses is provided by the case of American Indians and Eskimos. These groups are hyperresponsive to alcohol's effects (that is, they respond immediately and intensely to the alcohol in their systems) and yet have the highest alcoholism rates in the United States.
The claim of inheritance of alcoholism from the opposite theoretical direction??"that these groups succumb to alcoholism so readily because they metabolize alcohol so quickly??"likewise does not succeed. Groups that share the hypermetabolism of alcohol that Eskimos and Indians display (called Oriental flush), such as the Chinese and Japanese, have among the lowest alcoholism rates in America. The disjunctive connection between obvious metabolic characteristics and drinking habits actually contraindicates significant biological determinism in alcoholism (Mendelson and Mello 1979a).
The basic problem with genetic models of alcoholism is the absence of a reasonable link to the drinking behaviors in question. Why do any of the proposed genetic mechanisms lead people to become compulsive imbibers?
For example, in the case of an insensitivity to alcohol's effects, why wouldn't an individual who can't reliably detect that he has drunk too much simply learn from experience (in the absence of any proposed genetic compulsion to drink) to limit himself to a safer number of drinks? Do such drinkers simply choose to drink at unhealthy levels and to experience the extreme negative consequences of drinking that, after years, may lead to alcoholism (Vaillant 1983)? If so, why? That is the question.
On the other hand, the proposed differences in metabolizing alcohol and changes in brain functioning due to drinking are extremely subtle when compared with the gross effects of Oriental flush. Yet even groups characterized by Oriental flush, like the Indians and the Chinese, can show diametrically opposite responses to the same intense physiological changes. If a given individual did indeed have an extreme reaction to alcohol, why would he not become the type of drinker who announces, "I only have a drink or two because otherwise I become giddy and make a fool of myself"?
For those drinkers for whom alcohol might produce a desirable change in brain waves, why does the person prefer this state over others or other ways of gaining the same effect? The variation in behavior that is left unaccounted for in the most optimistic of these models is such as to discount the potential gain from the pursuit of as yet unestablished links between genetically inherited reactions to alcohol and alcoholic behavior. Finally, since all studies have found that it is sons and not daughters who most often inherit the risk of alcoholism (Cloninger et al. 1978), in what comprehensible ways can any of the genetic mechanisms thus far suggested for alcoholism be sex-linked?
The Endorphin-Deficiency Explanation of Narcotic Addiction
Since the primary assumption about narcotics has been that the drugs are equally and inevitably addictive for everyone, pharmacological theories of narcotic addiction have rarely stressed individual biological proclivities to be addicted. It was only a matter of time, however, before pharmacological and biological theorists began to hypothesize inherited mechanisms to account for differences in addictive susceptibility.
When Dole and Nyswander (1967) introduced the ideas that narcotic addiction was a "metabolic disease" and that the tendency to become addicted outlived the actual dependence on a drug, the way was opened to suggest that "metabolic disorder could precede as well as be precipitated by opiate use" (Goldstein, cited in Harding et al. 1980: 57). That is, not only might habitual narcotic use cause a chronic and residual need for drugs, but people conceivably might already have had such a need when they started taking drugs and came to rely on them.
The discovery that the body produces its own opiates, called endorphins, presented a plausible version of this mechanism. Endorphin theorists like Goldstein (1976b) and Snyder (1977) speculated that addicts may be characterized by an inbred endorphin deficiency that leaves them unusually sensitive to pain. Such people would then especially welcome??"and might even require??"the elevation of their pain threshold brought on by narcotics.
Heroin addicts have not yet been demonstrated to show unusual levels of endorphins. Moreover, this type of theorizing is badly strained??"as are all metabolic theories of addiction??"by the commonplace observations of drug abuse and addiction that were noted in chapter 1. Addicts do not in fact indicate a chronic, habitual need for narcotics. They regularly alter the type and amount of drug they use, sometimes abstaining or quitting altogether as they age.
Most of the Vietnam veterans who were addicted in Asia and who then used narcotics in the United States did not become readdicted. Noting that almost none of the patients introduced to a narcotic in the hospital indicate a prolonged desire for the drug, we may wonder why so small a percentage of the general population displays this endorphin deficiency.
Endorphin deficiency and other metabolic models suggest a course of progressive and irreversible reliance on narcotics that actually occurs in only exceptional and abnormal cases of addiction. Those with inbred metabolic defects could conceivably account for only a small percentage of those who become addicted over their lifetimes.
Why would the narcotic addiction that disappeared for most Vietnam veterans (or for the many other addicts who outgrow it) differ fundamentally from all other kinds of addiction, such as the kind that persists for some people? To accept this dichotomous view of addiction violates the basic principle of scientific parsimony, by which we should assume that the mechanisms at work in a large portion of cases are present in all cases.
This is the same error made by psychologists who concede (without empirical provocation) that some alcoholics may indeed have constitutional traits that cause them to be alcoholic from their first drink even as research shows all alcoholics to be responsive to situational rewards and to subjective beliefs and expectations.
In his influential internal-external model of obesity, Schachter (1968) proposed that fat people had a different style of eating, one that depended on external cues to tell them when to eat or not. Unlike those of normal weight, Schachter's overweight subjects apparently could not rely on internal physiological signs to decide whether they were hungry.
As a social psychologist, Schachter originally emphasized cognitive and environmental stimuli that encouraged the obese to eat. However, his model left open the question of the source of this insensitivity to somatic cues, suggesting the probability that this was an inherited trait. Schachter's (1971) view of the sources of overeating became increasingly physiological in nature when he began comparing the behavior of ventromedial-lesioned rats with obese humans.
Several of Schachter's prominent students followed his lead in this direction. For example, Rodin (1981) eventually rejected the internal-external model (as most researchers have by now) with an eye toward locating a neurological basis for overeating. Meanwhile Nisbett (1972), another Schachter student, proposed an extremely popular model of body weight based on an internal regulatory mechanism, called set-point, which is inherited or determined by prenatal or early childhood eating habits.
Peele (1983b) analyzed Schachter's evolution into a purely biological theoretician in terms of biases Schachter and his students had shown all along against personality dynamics; against group, social, and cultural mechanisms; and against the role of values and complex cognitions in the choice of behavior.
As a result, the Schachter group consistently failed to pick up discrepant indicators in their obesity research, some of which led eventually to the jettisoning of the internal-external model. For example Schachter (1968) noted that normal-weight subjects did not eat more when they were hungry (as predicted) because they found the type of food and the time of day inappropriate for eating.
In another study that had important implications, Nisbett (1968) discovered that formerly overweight subjects who were no longer obese behaved similarly to obese subjects in an eating experiment. That is, they ate more after having been forced to eat earlier than when they had not eaten before. Nisbett interpreted these results as showing that these subjects were unable to control their impulses to overeat and could therefore not be expected to keep excess weight off.
This line of thinking was solidified in Nisbett's set-point hypothesis, which held that the hypothalamus was set to defend a specific body weight and that going below this weight stimulated a greater desire to eat. The idea that obese people could not lose weight, based on laboratory studies and the performance of clients in weight-loss programs, had been the central tenet in all of the Schachter group's work on obesity (cf. Schachter and Rodin 1974; Rodin 1981).
Yet such pessimism seems an unlikely deduction from a study like Nisbett's (1968), in which subjects who had been obese and who continued to display an external eating style had indeed lost weight. When Schachter (1982) actually questioned people in the field about their weight-loss histories, he found remission was quite common in obesity: of all those interviewed who had ever been obese and who had tried to lose weight, 62.5 percent were currently at normal weight.
Schachter's serendipitous finding disputed the entire thrust of over a decade's research??"namely, that people were locked into obesity by biological forces. The idea would not die easily, however. Another Schachter student and his colleague recorded Schachter's (1982) finding but dismissed its significance by indicating it was probably only those obese subjects who were above their set-points who had been able to lose weight in this study (Polivy and Herman 1983: 195-96).
Polivy and Herman based this calculation on the estimate that from 60 to 70 percent of obese people were not obese in childhood. Their assertion requires that we believe that nearly all of the people in the Schachter study who have been overweight for reasons other than biological inheritance (and only these) had lost weight. Yet undoubtedly many in this category would remain fat for whatever presumably nonset-point reasons had caused them to become obese in the first place. Rather than being the underlying source of obesity its adherents had painted it to be, set-point now seemed not to be a major factor in most cases of overweight.
Polivy and Herman's (1983) description of their outlook did not reflect this understanding about set-point and obesity. Instead, they argued that "for the foreseeable future, we must resign ourselves to the fact that we have no reliable way to change the natural weight that an individual is blessed or cursed with" although "perhaps, as research progresses, we will be able to imagine such biological interventions??"including even genetic manipulations" that will enable people to lose weight (p. 52). Polivy and Herman furthermore attributed binge overeating??"the extreme of which is bulimia??"to people's attempts to restrain their eating in the effort to go below their natural weight (see chapter 5).
These researchers' work agrees with that of popular writers (Bennett and Gurin 1982) and the dominant research approaches in the field (Stunkard 1980) in maintaining a view of human eating and overeating that is essentially the same as that held by biological theorists of alcoholism and drug addiction toward drinking and drug consumption. In all cases, people are seen to be under the sway of invariant forces that, in the long run, they cannot hope to contravene.
Meanwhile, Garn and his coworkers (1979) have shown that similarities in weight levels among people who live together are a result of similar eating habits and energy expenditure. This "cohabitational effect" holds for husbands and wives and is the largest factor in weight similarities between parents and adopted offspring.
People who live together who become fat do so together (Garn et al. 1979). The longer parents and their children live together (even when the children are age 40) the more they resemble each other in fatness. The longer parents and children live separately, the less pronounced such similarities become until they approach 0 at the extremes of separation (Garn, LaVelle, and Pilkington 1984). Garn, Pilkington, and LaVelle (l984), observing 2,500 people over two decades, found "those . . . who were lean to begin with generally increased in fatness level. Those who were obese to begin with generally decreased in fatness level" (pp. 90-91). "Natural weight" may be a very variable thing, influenced by the same social values and personal coping strategies that affect all behavior (Peele 1984).
The enormity of the implications of the genetic transmission of addictive impulses is driven home by several theories claiming that people are compelled by chemical imbalances to form unhealthy, compulsive, and self-destructive interpersonal relationships. Tennov (1979) maintained that such "limerent" people, who are in every other way indistinguishable from other people, have a biological propensity to fall head-over-heels in love and create disastrous romantic attachments.
Liebowitz (1983) proposed that a failure in neurochemical regulation??"similar to that hypothesized to cause manic-depressive reactions leads people (almost exclusively women) to fall heatedly in love, often with inappropriate partners, and to become inordinately depressed when the relationships fail. These theories illustrate mainly the temptation to believe that compelling motivations must have a biological source and the desire to mechanize human differences, imperfections, and mysteries.
Global Biologic Theories of Addiction
Peele and Brodsky (1975), in the book Love and Addiction, also described interpersonal relationships as having addictive potential. The thrust of their version of interpersonal addiction, however, was exactly the opposite of that in Liebowitz (1983) and Tennov (1979): Peele and Brodsky's aim was to show that any powerful experience can form the object of an addiction for people predisposed by combinations of social and psychological factors.
Their approach was antireductionist and rejected the deterministic force of inbred, biological, or other factors outside the realm of human consciousness and experience. Their work signaled a burst of addiction theorizing in areas other than substance abuse, the bulk of which??"paradoxically??"sought to analyze these phenomena at a biological level. The result has been the proliferation of biologic theories to account both for the range of compulsive involvements people form and for the tendency some people show to be addicted to a host of substances.
Smith (1981), a medical clinician, has posited the existence of an "addictive disease" to account for why so many of those who become addicted to one substance have prior histories of addiction to dissimilar substances (cf. "The Collision of Prevention and Treatment" 1984). It is impossible to explain??"as Smith attempts to do??"how innate, predetermined reactions could cause the same person to become excessively involved with substances as disparate as cocaine, alcohol, and Valium.
In examining the generally strong positive correlations among tobacco, alcohol, and caffeine use, Istvan and Matarazzo (1984) explored the possibilities both that these substances are "linked by reciprocal activation mechanisms" and that they may be linked by their "pharmacologically antagonistic . . . effects" (p. 322). The evidence here is rather that substance abuse exceeds biological predictability. The fact of multiple addictions to myriad substances and nonsubstance-related involvements is primary evidence against genetic and biological interpretations of addiction.
Nonetheless, neuroscientists put forward biological theories of just this degree of universality. One researcher (Dunwiddie 1983: 17) noted that drugs of abuse such as opiates, amphetamine, and cocaine can pharmacologically stimulate many of the brain centers identified as reward centers.... On the other hand, there is considerable evidence that certain individuals have an enhanced liability for drug abuse, and frequently misuse a variety of seemingly unrelated drugs.
It is interesting to speculate that for various reasons, perhaps genetic, perhaps developmental or environmental, the normal inputs to these hypothetical "reward pathways" function inadequately in such individuals. If this were the case, there may be a biological defect underlying poly-drug abuse.
While piling hypothesis upon hypothesis, Dunwiddie's description presents no actual research findings about drug abusers, nor does it present a specific hypothetical link between deficient "reward pathways" and "polydrug abuse." It would seem the author thinks people who get less reward from drugs are more likely to abuse them.
Milkman and Sunderwirth's (1983) neurological model of addiction is not limited to drug abuse (as nothing in Dunwiddie's account would so limit it). These authors believe that addiction can result from any "self-induced changes in neurotransmission," where the more neurotransmitters that are involved "the faster the rate of firing," leading to the "elevated mood sought by cocaine users, for example" (p. 36).
This account is actually a social-psychological one masquerading as neurological explanation, in which the writers introduce social and psychological factors such as peer influence and low self-esteem into their analysis by suggesting "that the enzyme produced by a given gene might influence hormones and neurotransmitters in a way that contributes to the development of a personality potentially more susceptible to . . . peer group pressure" (p. 44).
Both Dunwiddie's and Milkman and Sunderwirth's analyses cloak experiential events in neurological terminology without reference to any actual research that connects biological functioning to addictive behavior. These models represent almost ritualistic conceptions of scientific enterprise, and while their analyses are caricatures of contemporary scientific model building, they come unfortunately close to mainstream assumptions about how the nature of addiction is to be interpreted.
Exposure Theories: Biological Models
The Inevitability of Narcotic Addiction
Alexander and Hadaway (1982) referred to the prevailing conception of narcotic addiction among both lay and scientific audiences??"that it is the inevitable consequence of regular narcotics use??"as the exposure orientation. So entrenched is this viewpoint that Berridge and Edwards (1981)??"while arguing that "Addiction is now defined as a disease because doctors have categorized it thus" (p. 150)??"refer readers to an appendix in which Griffith Edwards declared "anyone who takes an opiate for a long enough period of time and in sufficient dose will become addicted" (p. 278). This view contrasts with conventional beliefs about alcohol that would reject the same statement with the word "alcohol" substituted for "an opiate."
Underlying the exposure model is the assumption that the introduction of a narcotic into the body causes metabolic adjustments that require continued and increasing dosages of the drug in order to avoid withdrawal. No alteration in cell metabolism has yet been linked with addiction, however. The most prominent name in metabolic research and theory, Maurice Seevers, characterized efforts during the first sixty-five years of this century to create a model of addictive narcotic metabolism to be "exercises in semantics, or plain flights of imagination" (cited in Keller 1969: 5).
Dole and Nyswander (1967; cf. Dole 19